[comex]

Huh. To my eyes, the switch version is "obviously" much more clear! It's significantly more compact, with less visual noise you have to mentally ignore. And having the possible inputs and outputs lined up vertically (especially with nothing else between them) makes it easy to understand the logic. Plus, using a switch indicates at a glance that the variable being tested is the same in all cases, whereas with an if-else-if chain there could be any weird expression hiding in one of the conditions.

Eye of the beholder, indeed.

[vortico]

Perhaps a better comparison to the switch statement is

  if (c == 4) return 30;
  if (c == 5) return 70;
  return 0;

[comex]

Yep, that's a big improvement in my eyes, making it comparable from a readability perspective. In my opinion, that leaves more minor pros and cons of each approach, depending on the language. For example, in C and C++, as kazinator mentioned elsewhere in the thread, GCC and Clang now have an option to require all fallthroughs to be explicitly annotated, which effectively removes one of the main downsides of switch. And switch is better at detecting duplicate cases, as well as checking for exhaustivity if you're matching against enum variants. On the other hand, if you change the example so that the cases don't return (perhaps they assign to a variable instead), with switch you would need to add a `break;` after each one, increasing visual noise. If statements don't have that problem. Or if you change the example so that one of the values `c` is compared to is not a constant, some languages (C, C++, Java but not JavaScript, PHP) don't allow that with switch statements...